Nonconcordance of mutagenicity and carcinogenicity assays results when data from short-term mutation assays do not predict the results of 2 year bioassays. Our studies on those chemicals which produced positive mutagenicity in vitro and failed to produce carcinogenicity in NTP bioassays demonstrated the requirement for cell proliferation in the early stages of chemical exposure for positive carcinogenicity results. Chemicals which fail to cause cell proliferation also fail to cause carcinogenesis, regardless of their activity in mutagenesis assays. We have examined a number of chemicals to date and observed that the ability to cause cell proliferation and carcinogenesis is organ-specific and site-specific within an organ, and may be sex- and species-specific. We are also studying the mechanism(s) whereby chemicals that induce peroxisomes or cytochrome P450 isozymes produce hepatocarcinogenesis, and the relationship of this effect to human risk following exposure to these chemicals. Recent studies have used transgenic (Big Blue?) mice to detect in vivo mutagenesis induced by DAT isomers. In NTP rodent bioassays, 2,4-DAT is a potent hepatocarcinogen, whereas 2,6-DAT is not tumorigenic either in rats or mice even when administered at a higher dose. Mice were fed 2,4- or 2,6-DAT in the diet at 0 or 1000 ppm. However, the mutant frequency for animals treated with the carcinogen 2,4-DAT was significantly higher (p<0.01) than age-matched control animals or 2,6-DAT treated animals at 90 days. These results indicate that in vivo mutagenesis is induced only by the carcinogenic isomer of DAT and that assay duration is a significant variable for detecting mutant frequency and that cell proliferation may be requisite for expression of chemical-induced mutagenicity in vivo and ultimately for expression of carcinogenicity.